plan-design-review
Review implementation plans and specs for design quality before coding. Use when a plan needs UI/UX critique, missing state coverage, AI-slop detection, responsive/accessibility checks, or unresolved design decisions.
Best use case
plan-design-review is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Review implementation plans and specs for design quality before coding. Use when a plan needs UI/UX critique, missing state coverage, AI-slop detection, responsive/accessibility checks, or unresolved design decisions.
Teams using plan-design-review should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/plan-design-review/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How plan-design-review Compares
| Feature / Agent | plan-design-review | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Review implementation plans and specs for design quality before coding. Use when a plan needs UI/UX critique, missing state coverage, AI-slop detection, responsive/accessibility checks, or unresolved design decisions.
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
Related Guides
SKILL.md Source
# Plan Design Review Interactive design review for plans and specs, before any code is written. Rates each design dimension 0-10, explains what a 10 looks like, then edits the plan to close the gap. Use when: a plan or spec exists and you want to catch missing UI states, vague layouts, AI slop risk, and unresolved design decisions before implementation. Works on any plan file, not just ones from yc-office-hours. ## How It Works This is an interactive review. For each dimension, rate it, explain what excellence looks like, then fix the plan or ask the user to make a design choice. One issue at a time (don't dump a wall of findings). --- ## Initial Rating Read the plan. Give an honest overall design score (0-10) with a one-paragraph explanation of why. Be blunt. Example: > **Initial Design Rating: 4/10** > This plan describes a user dashboard but never specifies what the user sees first. It says "cards with icons," which looks like every SaaS template. It mentions zero loading states, zero empty states, and no mobile behavior. --- ## 7 Review Passes Work through each pass in order. For each: 1. Rate the dimension 0-10 2. Describe what a 10 looks like for this specific plan 3. Fix obvious gaps directly in the plan 4. For genuine tradeoffs, ask the user to decide (one question at a time) ### Pass 1: Information Architecture - Is there a clear content hierarchy (primary / secondary / tertiary) for every screen? - What's the most important thing on each page? Is it visually dominant? - Does the navigation structure match how users think about the product? ### Pass 2: Interaction State Coverage For every UI feature, check all 5 states: - **Empty** — what does the user see before any data exists? - **Loading** — what happens while data is being fetched? - **Partial** — what if there's some data but not all? - **Error** — what does the user see when something breaks? - **Ideal** — what does the full, working state look like? Count them: `[number of features] × 5 = [total states needed]`. How many does the plan specify? Report the gap. ### Pass 3: User Journey - What's the first thing a new user sees? - What's the path from landing to core value? - Where can a user get stuck or confused? - Are there dead ends (pages with no clear next action)? ### Pass 4: AI Slop Risk Flag patterns that scream "AI-generated template": - "Clean, modern UI with cards and icons" - Hero sections with gradients - 3-column icon grids - Uniform border-radius on everything - Generic stock-photo-style imagery descriptions - "Sleek" / "intuitive" / "seamless" as design specs For each one found, propose a specific, intentional alternative. ### Pass 5: Design System Alignment - Does the plan reference an existing design system? If yes, does it follow it? - If no design system exists, does the plan at least specify: colors, typography, spacing, component patterns? - Are UI descriptions specific enough to implement without guessing? ### Pass 6: Responsive & Accessibility - Does the plan specify behavior at mobile, tablet, and desktop breakpoints? - Are touch targets large enough (44px minimum)? - Is there sufficient color contrast? - Does the plan mention keyboard navigation, screen readers, or focus management? - Does `prefers-reduced-motion` need consideration? ### Pass 7: Unresolved Design Decisions - Are there any "TBD" or vague sections in the plan? - Are there places where two reasonable approaches exist and nobody picked one? - List every unresolved decision. Ask the user to resolve each one. --- ## Scoring & Re-runs After all passes, give a final score. ``` Pass Before After Information Architecture 3/10 8/10 Interaction State Coverage 2/10 7/10 User Journey 5/10 8/10 AI Slop Risk 4/10 9/10 Design System Alignment 6/10 7/10 Responsive & Accessibility 3/10 6/10 Unresolved Decisions 1/10 8/10 ───────────────────────────────────────── Overall 4/10 7/10 ``` On re-runs, passes already at 8+ get a quick confirmation. Passes below 8 get full treatment. --- ## Conversation Style - One finding at a time. Don't overwhelm with a list of 20 problems. - Fix the obvious stuff silently. Only ask about genuine tradeoffs. - Be specific. "The empty state needs work" is useless. "When a user has zero transactions, they see a blank white page with no guidance on what to do next" is actionable. - Rate honestly. A 4 is a 4. Don't give 7s to be nice. - When describing what a 10 looks like, reference real products when possible. ## Anti-Patterns - Don't redesign the product. You're reviewing the plan's design completeness, not proposing a different product. - Don't add features disguised as design feedback. - Don't skip passes because "it's fine." Rate it and move on quickly if it is. - Don't give generic feedback that applies to any plan. Every comment should reference something specific in this plan.
Related Skills
plan
Plan mode for Hermes Agent — inspect context, write a markdown plan into the active workspace's `.hermes/plans/` directory, and do not execute the work.
frontend-design
Create distinctive, production-grade frontend interfaces with high design quality. Generates creative, polished code that avoids generic AI aesthetics. Use when the user asks to build web components, pages, artifacts, posters, or applications, or when any design skill requires project context.
design-review
Run a product design review on a feature or site. Answers 13 design questions, runs Nielsen Norman heuristic evaluation, builds before/after visual fixes, and deploys a shareable report to Surge. Use when asked to review a design, audit UX, do a design review, or analyze a product's user experience.
design-mode
Complement to frontend-design — embodies Anthropic's design-surface system prompt (the "expert designer" persona that produces HTML artifacts, decks, prototypes, and animations in a filesystem-based project). Use when the user wants to enter "design mode" and operate as a design-artifact generator with the full Anthropic design workflow (asking questions first, exploring design systems, Tweaks, speaker notes, starter components, verifier agent), or when frontend-design has set aesthetic direction and the next step is producing the artifact. Trigger phrases: "design mode", "design artifact", "make a deck", "hi-fi prototype", "Anthropic design mode", "Claude design sys prompt", "design like claude.ai/design".
design-md
Use when creating, editing, or validating DESIGN.md files, design tokens, DTCG token specs, Tailwind theme exports, design system documentation, WCAG contrast checks, or Google Stitch-compatible design specs.
claude-design
Design one-off HTML artifacts (landing, deck, prototype).
writer
Write content in Eric's voice — articles, blog posts, tweets, social media posts, marketing copy, newsletter drafts. Loads WRITING-STYLE.md and enforces kill phrases.
positioning-angles
Use when defining product positioning, choosing strategic angles, crafting value propositions, competitive positioning, product messaging, differentiation strategy, or go-to-market angles. Also use for 'how should I position my app', 'what angle should I use', 'painkiller vs vitamin', or 'market positioning'.
outline-generator
Use when generating outlines, article structures, content outlines, blog outlines, planning article sections, structuring posts, breaking down topics into sections, or organizing ideas for long-form content. Also use for 'outline this', 'structure this article', or 'plan the sections'.
last30days-open
Use only when the user explicitly asks for the open variant of last30days, including watchlists, briefings, and history queries. Sources: Reddit, X, YouTube, web.
last30days
Use when researching what happened in the last 30 days on a topic. Also triggered by 'last30'. Sources: Reddit, X, YouTube, web. Produces expert-level summary with copy-paste-ready prompts.
hooks
Use when generating hooks, headlines, titles, and scroll-stopping openers for content. Also use when analyzing viral posts, Reels, TikToks, YouTube Shorts, or successful social examples to extract reusable hook patterns and improve hook guidance.